Project Summary /Abstract: The HMO Research Network's (HMORN) Diabetes Multi-Center Research Consortium (DMCRC) Coordinating Center proposes to build a national research network and a multi-system distributed database for conducting comparative effectiveness research (CER) in the treatment and prevention of diabetes mellitus. The network is comprised of 31 investigators from 12 integrated healthcare delivery systems and academic institutions who study diabetes mellitus, gestational diabetes mellitus, childhood and adult obesity, weight management, medical informatics and biostatistics. The network has broad experience conducting system-based interventions to improve treatment of diabetes or to support lifestyle changes to prevent diabetes, as well as broad experience conducting CER. The 12-system database will capture and standardize detailed demographic and longitudinal electronic health record (EHR)-derived clinical data on more than 750,000 persons with diabetes and a much larger membership without diabetes. Once constructed, the database will be used to conduct and publish surveillance data on trends (2005- 2012) in the incidence, prevalence, treatment and outcomes of diabetes in this population. It will also be used to conduct two CER studies. The first is a cluster randomized, multi-system intervention that will rapidly identify diabetes patients with "early nonadherence" to newly prescribed anti-diabetic, antihypertensive, and lipid-lowering medications. Early nonadherence refers to patient failure to fill either the first prescription of a new medication or the first refill, and has been shown to be more frequent than "ongoing" nonadherence. This information will be provided to population management outreach teams in 3-5 health systems via the EHR or other electronic database, along with training in counseling patients on adherence problems. Study outcomes of subsequent risk factor change and adherence to the new medication will be compared within each system to those for patients on teams that do not receive the new information. The second CER study is an observational evaluation of various communication, counseling and referral strategies as provided to women who have had an abnormal glucose test (fasting glucose or oral glucose tolerance test) in the first 6 months after a pregnancy complicated by gestational diabetes mellitus (GDM) and who are therefore at very high risk for developing Type 2 diabetes. Study outcomes include weight change over one year and self-reported physical activity four months after the abnormal test. In each study, we will collect process of care data by applying natural language processing (NLP) to clinical text in EHRs and survey patients for information on demographic, clinical, and behavioral variables. These NLP-derived and patient-reported variables will be studied as possible mediators of treatment effectiveness or as potential modifiers of effectiveness (i.e. variables that identify patient subgroups that benefit less from usual interventions. Variables that prove to be important as mediators or modifiers may be good candidates for routine collection and incorporation into EHR's and/or future diabetes registries. PUBLIC HEALTH RELEVANCE: A network of 32 health services and intervention scientists affiliated with the HMO Research Network propose to build a national database for conducting comparative effectiveness research on treatment and on approaches to preventing diabetes mellitus. The database will include information on more than 750,000 persons with diabetes. It will be used to generate ongoing information on the occurrence, treatments and outcomes of diabetes and to conduct 2 studies of approaches to treating and preventing this condition.